Live Event Content · Disruptive AI technologies transforming healthcare and medical education
Video34 min·Published Aug 2023Older

Disruptive AI technologies transforming healthcare and medical education

With Dr. Rami Shaman · hosted by Dr. Todd Ponsky & Dr. Em Gootee
Try
Intelligent Search· scoped to cancer · not medical adviceSearch the whole library →
What the experts said9 expert statements · 24 host summaries
Netflix wanted to hire an AI specialist for a million dollars a year salary
EpidemiologicalRami Shaman
AI is replacing people who don't know how to use AI with people who know how to use AI
OpinionRami Shaman
ChatGPT is much better than Google for information retrieval
OpinionTodd Ponsky
Todd Ponsky uses ChatGPT 15 times a day
EpidemiologicalTodd Ponsky
Todd Ponsky has not personally written a letter of recommendation in the last 3 months, using ChatGPT instead by uploading CVs and personal statements
EpidemiologicalTodd Ponsky
Free online filters exist that can detect with a certain amount of certainty whether text is AI-created
Clinical
Colleges know that students are using ChatGPT to write their college essays
EpidemiologicalTodd Ponsky
Cincinnati Children's Hospital can now create videos of every faculty member teaching a topic using deep fake technology with their voice and likeness, without the faculty member doing anything
ClinicalTodd Ponsky
Cincinnati Children's has a membership for Wysa mental health chatbot that employees can access with their work email address
ClinicalRami Shaman
Healthcare produces massive amounts of data through patient records, lab results, imaging, and more, which when digitized become the foundation of AI systems
Host summaryEm Gootee summarizing the discussion · not cited in answers
AI algorithms can analyze radiology images to detect early signs of cancer or predict a patient's risk of developing certain diseases based on medical history
Host summaryRami Shaman summarizing the discussion · not cited in answers
AI systems continually enhance effectiveness by feeding them more data and refining algorithms based on prediction accuracy feedback
Host summaryEm Gootee summarizing the discussion · not cited in answers
AI can help with earlier disease detection, perfecting robotic surgery, administrative tasks, reducing human errors through diagnostic assistance, and personalized treatment recommendations based on genetic makeup and disease history
Host summaryRami Shaman summarizing the discussion · not cited in answers
Google's DeepMind AlphaFold system predicts protein 3D structure from amino acid sequence with potential applications in understanding diseases and drug discovery
Host summaryRami Shaman summarizing the discussion · not cited in answers
In the last 60 years, the scientific community has determined the structures of over 180,000 proteins in atomic detail using advanced experimental methods
Host summaryEm Gootee summarizing the discussion · not cited in answers
Determining the exact structure of a protein remains an expensive and time-consuming process, meaning we only know the exact 3D structure of a tiny fraction of proteins known to science
Host summaryEm Gootee summarizing the discussion · not cited in answers
AlphaFold Database provides open access to over 200 million structure predictions to accelerate scientific research
Host summaryRami Shaman summarizing the discussion · not cited in answers
Machine learning algorithms can be trained on vast datasets of images to identify and diagnose diseases, detecting tumors in X-rays or MRIs often with greater accuracy than human experts
Host summaryEm Gootee summarizing the discussion · not cited in answers
Zebra Medical Vision uses ML to analyze medical imaging and detect diseases in organs such as liver, cardiovascular, and lungs
Host summaryRami Shaman summarizing the discussion · not cited in answers
Path AI concentrates on delivering more accurate diagnosis from pathology slides
Host summaryRami Shaman summarizing the discussion · not cited in answers
Machine learning is instrumental in predictive analytics such as patient readmissions, potential outbreaks, patient no-shows, and adverse events
Host summaryEm Gootee summarizing the discussion · not cited in answers
ML algorithms, particularly deep learning models, can process and analyze vast amounts of text including doctors' notes and extract valuable information
Host summaryEm Gootee summarizing the discussion · not cited in answers
23andMe employs machine learning to analyze genetic data and deliver personalized health insights
Host summaryEm Gootee summarizing the discussion · not cited in answers
Tempus leverages ML to assist physicians in making more personalized treatment decisions, particularly for oncology patients
Host summaryRami Shaman summarizing the discussion · not cited in answers
ML-powered chatbots and applications can provide 24/7 support to individuals experiencing stress, anxiety, depression, or other mental health concerns, offering coping strategies and resources as an interim solution when human therapists are not immediately available
Host summaryEm Gootee summarizing the discussion · not cited in answers
Wysa provides emotional help by tracking mood, highlighting optimism, and reframing thinking in an approachable manner
Host summaryRami Shaman summarizing the discussion · not cited in answers
Large language models can aid healthcare professionals by providing real-time evidence-based responses to clinical inquiries, scanning immense volumes of medical literature in seconds, summarizing findings, or suggesting diagnoses based on symptom descriptions
Host summaryEm Gootee summarizing the discussion · not cited in answers
LLMs can automate documentation, assist in medical data entry, and respond to routine queries, freeing up valuable time for medical professionals
Host summaryRami Shaman summarizing the discussion · not cited in answers
Glass AI leverages GPT-4 combined with evidence-based guidelines, schemas, and case studies created by clinicians to generate differential diagnoses or draft clinical plans based on submitted clinical problem representations
Host summaryEm Gootee summarizing the discussion · not cited in answers
To use ChatGPT effectively, you must give it a role (e.g., 'you are a pediatric physician') before asking questions, a technique called prompt engineering
Host summaryEm Gootee summarizing the discussion · not cited in answers
Microsoft Co-Pilot can access a folder of presentations and papers to write content based solely on the user's own work without taking anything from outside sources
Host summaryEm Gootee summarizing the discussion · not cited in answers
Files uploaded to ChatGPT are completely secure with end-to-end encryption and do not go into the internet or cloud
Host summaryEm Gootee summarizing the discussion · not cited in answers
Microsoft Co-Pilot will be included in the next version of Microsoft Office within a couple of months
Host summaryEm Gootee summarizing the discussion · not cited in answers